语言大模型 / LLM - arXivDaily 专题

2509.26497 2025-10-01 cs.CV 92%

Revealing the Power of Post-Training for Small Language Models via Knowledge Distillation

Miao Rang, Zhenni Bi, Hang Zhou, Hanting Chen, An Xiao, Tianyu Guo, Kai Han, Xinghao Chen, Yunhe Wang

专题命中效率与部署：language model(title,abstract)；post-training(title,abstract)；small language model(title)；large language model(abstract)

Comments 7

2507.14017 2025-10-01 cs.CL cs.LG 90%

Efficient Temporal Tokenization for Mobility Prediction with Large Language Models

Haoyu He, Haozheng Luo, Yan Chen, Qi R. Wang

机构 * Northeastern University, Boston, MA（东北大学）； Northwestern University, Evanston, IL（西北大学）

专题命中效率与部署：large language model(title,abstract)；language model(title,abstract)；LLM(abstract)；分类 cs.CL、cs.LG

Journal ref Proceedings of the 3rd Workshop on Efficient Systems for Foundation Models (ES-FoMo III) at ICML 2025

2509.25996 2025-10-01 cs.LG cs.CL 90%

CAST: Continuous and Differentiable Semi-Structured Sparsity-Aware Training for Large Language Models

Weiyu Huang, Yuezhou Hu, Jun Zhu, Jianfei Chen

机构 * Department of Computer Science and Technology, Institute for AI, BNRist Center, THBI Lab, Tsinghua-Bosch Joint ML Center, Tsinghua University（计算机科学与技术系、人工智能研究院、BNRist中心、THBI实验室、清华大学-博世联合机器学习中心、清华大学）

专题命中效率与部署：large language model(title,abstract)；language model(title,abstract)；pretraining(abstract)；分类 cs.CL、cs.LG

Comments Submitted to IEEE TPAMI

2509.24218 2025-10-01 cs.LG cs.AI 90%

Conda: Column-Normalized Adam for Training Large Language Models Faster

Junjie Wang, Pan Zhou, Yiming Dong, Huan Li, Jia Li, Xun Zhou, Qicheng Lao, Cong Fang, Zhouchen Lin

机构 * State Key Lab of General AI, School of Intelligence Science and Technology, Peking University（通用人工智能国家重点实验室，智能科学与技术学院，北京大学）； Singapore Management University（新加坡国立大学）； Nankai University（南开大学）； Beijing Normal University（北京师范大学）； ByteDance Seed（字节跳动种子）； Beijing University of Posts and Telecommunications（北京邮电大学）； Institute for Artificial Intelligence, Peking University（人工智能研究院，北京大学）； Pazhou Laboratory (Huangpu), Guangzhou, Guangdong, China（黄埔实验室（广州），广东，中国）

专题命中效率与部署：large language model(title,abstract)；language model(title,abstract)；LLM(abstract)；分类 cs.AI、cs.LG

2501.17391 2025-10-01 cs.CV cs.AI cs.CL 90%

LFTR: Learning-Free Token Reduction for Multimodal Large Language Models

Zihui Zhao, Yingxin Li, Yang Li

专题命中效率与部署：large language model(title,abstract)；language model(title,abstract)；post-training(abstract)；分类 cs.CL、cs.AI

2509.26593 2025-10-01 cs.HC 89%

Exploring Large Language Model as an Interactive Sports Coach: Lessons from a Single-Subject Half Marathon Preparation

Kichang Lee

专题命中效率与部署：large language model(title,abstract)；language model(title,abstract)；LLM(abstract)

Comments 23 pages, 21 figures

2506.04203 2025-10-01 cs.DC 89%

Cascadia: An Efficient Cascade Serving System for Large Language Models

Youhe Jiang, Fangcheng Fu, Wanru Zhao, Stephan Rabanser, Jintao Zhang, Nicholas D. Lane, Binhang Yuan

专题命中效率与部署：large language model(title,abstract)；language model(title,abstract)；LLM(abstract)

2509.11662 2025-10-01 cs.CV cs.AI cs.CL eess.IV 88%

MindVL: Towards Efficient and Effective Training of Multimodal Large Language Models on Ascend NPUs

Feilong Chen, Yijiang Liu, Yi Huang, Hao Wang, Miren Tian, Ya-Qi Yu, Minghui Liao, Jihao Wu

机构 * Huawei Technologies Co., Ltd.（华为技术有限公司）

专题命中效率与部署：large language model(title,abstract)；language model(title,abstract)；分类 cs.CL、cs.AI

2509.26482 2025-10-01 cs.AI 88%

TVS Sidekick: Challenges and Practical Insights from Deploying Large Language Models in the Enterprise

Paula Reyero Lobo, Kevin Johnson, Bill Buchanan, Matthew Shardlow, Ashley Williams, Samuel Attwood

机构 * TVS Supply Chain Solutions（TVS供应链解决方案）； Manchester Metropolitan University（曼彻斯特 Metropolitan 大学）

专题命中效率与部署：large language model(title,abstract)；language model(title,abstract)；分类 cs.AI

Comments Accepted at EthicalLLMs@RANLP2025

2505.20873 2025-10-01 cs.CV 88%

Fork-Merge Decoding: Enhancing Multimodal Understanding in Audio-Visual Large Language Models

Chaeyoung Jung, Youngjoon Jang, Jongmin Choi, Joon Son Chung

机构 * Korea Advanced Institute of Science and Technology (KAIST)（韩国科学技术院）

专题命中效率与部署：large language model(title,abstract)；language model(title,abstract)

2509.26200 2025-10-01 cs.NI cs.AI 85%

Toward an Unbiased Collective Memory for Efficient LLM-Based Agentic 6G Cross-Domain Management

Hatim Chergui, Miguel Catalan Cid, Pouria Sayyad Khodashenas, Daniel Camps Mur, Christos Verikoukis

机构 * i2CAT Foundation（i2CAT基金会）； ISI/ATH ； University of Patras（帕特拉大学）

专题命中效率与部署：LLM(title,abstract)；large language model(abstract)；language model(abstract)；分类 cs.AI

Comments 12 pages, 8 figures

2509.25919 2025-10-01 cs.DC cs.AI 85%

Accelerating LLM Inference with Precomputed Query Storage

Jay H. Park, Youngju Cho, Choungsol Lee, Moonwook Oh, Euiseong Seo

专题命中效率与部署：LLM(title,abstract)；large language model(abstract)；language model(abstract)；分类 cs.AI

2503.14043 2025-10-01 cs.LG 85%

Beyond Next Token Probabilities: Learnable, Fast Detection of Hallucinations and Data Contamination on LLM Output Distributions

Guy Bar-Shalom, Fabrizio Frasca, Derek Lim, Yoav Gelberg, Yftah Ziser, Ran El-Yaniv, Gal Chechik, Haggai Maron

专题命中效率与部署：LLM(title,abstract)；large language model(abstract)；language model(abstract)；分类 cs.LG

2408.00741 2025-10-01 cs.AI cs.AR cs.DC 85%

DynamoLLM: Designing LLM Inference Clusters for Performance and Energy Efficiency

Jovan Stojkovic, Chaojie Zhang, Íñigo Goiri, Josep Torrellas, Esha Choukse

机构 * University of Illinois at Urbana-Champaign（伊利诺伊大学厄巴纳-香槟分校）； Microsoft Azure Research - Systems（微软Azure研究-系统）

专题命中效率与部署：LLM(title,abstract)；large language model(abstract)；language model(abstract)；分类 cs.AI

2509.25803 2025-10-01 cs.IR cs.AI cs.CE cs.LG 85%

Better with Less: Small Proprietary Models Surpass Large Language Models in Financial Transaction Understanding

Wanying Ding, Savinay Narendra, Xiran Shi, Adwait Ratnaparkhi, Chengrui Yang, Nikoo Sabzevar, Ziyan Yin

机构 * JPMorgan Chase & Co.（摩根大通公司）

专题命中效率与部署：large language model(title)；language model(title)；分类 cs.AI、cs.LG

Comments 9 pages, 5 figures

2509.26182 2025-10-01 cs.DC 85%

Parallax: Efficient LLM Inference Service over Decentralized Environment

Chris Tong, Youhe Jiang, Gufeng Chen, Tianyi Zhao, Sibian Lu, Wenjie Qu, Eric Yang, Lynn Ai, Binhang Yuan

专题命中效率与部署：LLM(title,abstract)；large language model(abstract)；language model(abstract)

2509.25853 2025-10-01 cs.AR 85%

SAIL: SRAM-Accelerated LLM Inference System with Lookup-Table-based GEMV

Jingyao Zhang, Jaewoo Park, Jongeun Lee, Elaheh Sadredini

专题命中效率与部署：LLM(title,abstract)；large language model(abstract)；language model(abstract)

2405.18628 2025-10-01 cs.LG cs.CL 84%

Hardware-Aware Parallel Prompt Decoding for Memory-Efficient Acceleration of LLM Inference

Hao Mark Chen, Wayne Luk, Ka Fai Cedric Yiu, Rui Li, Konstantin Mishchenko, Stylianos I. Venieris, Hongxiang Fan

机构 * Imperial College London, UK（伦敦帝国学院）； Hong Kong Polytechnic University, Hong Kong（香港理工大学）； Samsung AI Center, Cambridge, UK（三星人工智能中心）

专题命中效率与部署：LLM(title)；large language model(abstract)；language model(abstract)；分类 cs.CL、cs.LG

Comments Accepted at EMNLP 2025. The code for this implementation is available at https://github.com/hmarkc/parallel-prompt-decoding

2509.25267 2025-10-01 cs.LG cs.AI cs.CL 83%

Dynamic Policy Induction for Adaptive Prompt Optimization: Bridging the Efficiency-Accuracy Gap via Lightweight Reinforcement Learning

Jiexi Xu

机构 * University of California, Irvine School of Information & Computer Science（加州大学尔湾分校信息与计算机科学学院）

专题命中效率与部署：LLM(abstract)；large language model(abstract)；language model(abstract)；prompting(abstract)

Comments 13 pages, 2 figures, 2 tables

2509.26328 2025-10-01 cs.CL 83%

Fast-dLLM v2: Efficient Block-Diffusion LLM

Chengyue Wu, Hao Zhang, Shuchen Xue, Shizhe Diao, Yonggan Fu, Zhijian Liu, Pavlo Molchanov, Ping Luo, Song Han, Enze Xie

机构 * The University of Hong Kong（香港大学）； NVIDIA（英伟达）； MIT（麻省理工学院）

专题命中效率与部署：LLM(title)；large language model(abstract)；language model(abstract)；分类 cs.CL

2509.26246 2025-10-01 cs.AI 83%

SlimPack: Fine-Grained Asymmetric Packing for Balanced and Efficient Variable-Length LLM Training

Yuliang Liu, Guohao Wu, Shenglong Zhang, Wei Zhang, Qianchao Zhu, Zhouyang Li, Chenyu Wang

机构 * Kling Infra, Kuaishou Technology（Kling Infra，快手科技）

专题命中效率与部署：LLM(title)；large language model(abstract)；language model(abstract)；分类 cs.AI

2509.25238 2025-10-01 cs.LG cs.AI 81%

PALADIN: Self-Correcting Language Model Agents to Cure Tool-Failure Cases

Sri Vatsa Vuddanti, Aarav Shah, Satwik Kumar Chittiprolu, Tony Song, Sunishchal Dev, Kevin Zhu, Maheep Chaudhary

专题命中效率与部署：language model(title)；language agent(abstract)；分类 cs.AI、cs.LG

2505.07675 2025-10-01 cs.LG cs.AI cs.CV 81%

Simple yet Effective Semi-supervised Knowledge Distillation from Vision-Language Models via Dual-Head Optimization

Seongjae Kang, Dong Bok Lee, Hyungjoon Jang, Sung Ju Hwang

机构 * VUNO Inc.（VUNO公司）； KAIST（韩国科学技术院）

专题命中效率与部署：language model(title,abstract)；分类 cs.AI、cs.LG

Comments 38 pages, 17 figures, preprint

2509.26124 2025-10-01 cs.CL 79%

Vocabulary Customization for Efficient Domain-Specific LLM Deployment

Christian Herold, Michael Kozielski, Nicholas Santavas, Yannick Versley, Shahram Khadivi

机构 * eBay Inc（eBay公司）

专题命中效率与部署：LLM(title,abstract)；分类 cs.CL

Comments Accepted at NEURIPS 2025 CCFM Workshop

2412.07448 2025-10-01 cs.AI 79%

Efficient Dynamic Ensembling for Multiple LLM Experts

Jinwu Hu, Yufeng Wang, Shuhai Zhang, Kai Zhou, Guohao Chen, Yu Hu, Bin Xiao, Mingkui Tan

机构 * South China University of Technology（华南理工大学）； Pazhou Laboratory（琶洲实验室）； Peng Cheng Laboratory（鹏城实验室）； Hong Kong Polytechnic University（香港理工大学）； Chongqing University of Posts and Telecommunications（重庆邮电大学）； Key Laboratory of Big Data and Intelligent Robot, Ministry of Education（教育部大数据与智能机器人重点实验室）

专题命中效率与部署：LLM(title,abstract)；分类 cs.AI

Comments IJCAI 2025

2509.26436 2025-10-01 cs.CV 78%

Post-Training Quantization via Residual Truncation and Zero Suppression for Diffusion Models

Donghoon Kim, Dongyoung Lee, Ik Joon Chang, Sung-Ho Bae

机构 * Department of Artificial Intelligence（人工智能系）； Kyung Hee University（庆尚大学）； Department of Electrical Engineering（电气工程系）； Department of Computer Science（计算机科学系）

专题命中效率与部署：post-training(title,abstract)

2509.25808 2025-10-01 cs.LG 77%

Improving Sampling Efficiency in RLVR through Adaptive Rollout and Response Reuse

Yuheng Zhang, Wenlin Yao, Changlong Yu, Yao Liu, Qingyu Yin, Bing Yin, Hyokun Yun, Lihong Li

机构 * University of Illinois Urbana-Champaign（伊利诺伊大学厄巴纳-香槟分校）； Amazon（亚马逊）

专题命中效率与部署：large language model(abstract)；language model(abstract)；post-training(abstract)；分类 cs.LG

2509.25712 2025-10-01 cs.LG 77%

Expert Merging: Model Merging with Unsupervised Expert Alignment and Importance-Guided Layer Chunking

Dengming Zhang, Xiaowen Ma, Zhenliang Ni, Zhenkai Wu, Han Shu, Xin Jiang, Xinghao Chen

机构 * Zhejiang University（浙江大学）； Huawei Noah’s Ark Lab（华为诺亚实验室）

专题命中效率与部署：LLM(abstract)；large language model(abstract)；language model(abstract)；分类 cs.LG

2509.22502 2025-10-01 cs.AI cs.HC 77%

InfiAgent: Self-Evolving Pyramid Agent Framework for Infinite Scenarios

Chenglin Yu, Yang Yu, Songmiao Wang, Yucheng Wang, Yifan Yang, Jinjia Li, Ming Li, Hongxia Yang

机构 * The Hong Kong University（香港大学）； The Hong Kong Polytechnic University（香港理工大学）

专题命中效率与部署：LLM(abstract)；large language model(abstract)；language model(abstract)；分类 cs.AI

Comments 9 pages of main content and 32 pages of others, 2 figures, under review as a conference paper at ICLR 2026

2509.25709 2025-10-01 econ.EM 75%

Leveraging LLMs to Improve Experimental Design: A Generative Stratification Approach

George Gui, Seungwoo Kim

专题命中效率与部署：LLM(abstract)；large language model(abstract)；language model(abstract)

1. 效率与部署 52 篇

Revealing the Power of Post-Training for Small Language Models via Knowledge Distillation

Efficient Temporal Tokenization for Mobility Prediction with Large Language Models

CAST: Continuous and Differentiable Semi-Structured Sparsity-Aware Training for Large Language Models

Conda: Column-Normalized Adam for Training Large Language Models Faster

LFTR: Learning-Free Token Reduction for Multimodal Large Language Models

Exploring Large Language Model as an Interactive Sports Coach: Lessons from a Single-Subject Half Marathon Preparation

Cascadia: An Efficient Cascade Serving System for Large Language Models

MindVL: Towards Efficient and Effective Training of Multimodal Large Language Models on Ascend NPUs

TVS Sidekick: Challenges and Practical Insights from Deploying Large Language Models in the Enterprise

Fork-Merge Decoding: Enhancing Multimodal Understanding in Audio-Visual Large Language Models

Toward an Unbiased Collective Memory for Efficient LLM-Based Agentic 6G Cross-Domain Management

Accelerating LLM Inference with Precomputed Query Storage

Beyond Next Token Probabilities: Learnable, Fast Detection of Hallucinations and Data Contamination on LLM Output Distributions

DynamoLLM: Designing LLM Inference Clusters for Performance and Energy Efficiency

Better with Less: Small Proprietary Models Surpass Large Language Models in Financial Transaction Understanding

Parallax: Efficient LLM Inference Service over Decentralized Environment

SAIL: SRAM-Accelerated LLM Inference System with Lookup-Table-based GEMV

Hardware-Aware Parallel Prompt Decoding for Memory-Efficient Acceleration of LLM Inference

Dynamic Policy Induction for Adaptive Prompt Optimization: Bridging the Efficiency-Accuracy Gap via Lightweight Reinforcement Learning

Fast-dLLM v2: Efficient Block-Diffusion LLM

SlimPack: Fine-Grained Asymmetric Packing for Balanced and Efficient Variable-Length LLM Training

PALADIN: Self-Correcting Language Model Agents to Cure Tool-Failure Cases

Simple yet Effective Semi-supervised Knowledge Distillation from Vision-Language Models via Dual-Head Optimization

Vocabulary Customization for Efficient Domain-Specific LLM Deployment

Efficient Dynamic Ensembling for Multiple LLM Experts

Post-Training Quantization via Residual Truncation and Zero Suppression for Diffusion Models

Improving Sampling Efficiency in RLVR through Adaptive Rollout and Response Reuse

Expert Merging: Model Merging with Unsupervised Expert Alignment and Importance-Guided Layer Chunking

InfiAgent: Self-Evolving Pyramid Agent Framework for Infinite Scenarios

Leveraging LLMs to Improve Experimental Design: A Generative Stratification Approach